Message ID | 1271602258.27235.200.camel@lb-tlvb-vladz |
---|---|
State | Changes Requested, archived |
Delegated to: | David Miller |
Headers | show |
On Sun, 2010-04-18 at 17:50 +0300, Vladislav Zolotarov wrote: > LRO cannot be used on virtualized environment, so it is best > to disable it on compile time in XEN kernel. [...] If you want to cripple your performance, I'm not going to stand in your way, but be aware that most distribution kernels now enable CONFIG_XEN. Ben.
From: Ben Hutchings <bhutchings@solarflare.com> Date: Sun, 18 Apr 2010 16:11:25 +0100 > On Sun, 2010-04-18 at 17:50 +0300, Vladislav Zolotarov wrote: >> LRO cannot be used on virtualized environment, so it is best >> to disable it on compile time in XEN kernel. > [...] > > If you want to cripple your performance, I'm not going to stand in your > way, but be aware that most distribution kernels now enable CONFIG_XEN. There is also zero reason to do this. When the XEN code or whatever setups up the bridge, LRO will be disabled by the bridge code. It automatically issues ethtool LRO disable commands to the devices involved in the bridge configuration. And it has done this for a _long_ time. The ipv4 routing code will do the same. This change is bogus and likely based upon behavior in some ancient distribution kernel that lacks the ethtool facilities. I absolutely refuse to apply this. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On Sun, 18 Apr 2010 20:58:58 -0700 (PDT) David Miller <davem@davemloft.net> wrote: > From: Ben Hutchings <bhutchings@solarflare.com> > Date: Sun, 18 Apr 2010 16:11:25 +0100 > > > On Sun, 2010-04-18 at 17:50 +0300, Vladislav Zolotarov wrote: > >> LRO cannot be used on virtualized environment, so it is best > >> to disable it on compile time in XEN kernel. > > [...] > > > > If you want to cripple your performance, I'm not going to stand in your > > way, but be aware that most distribution kernels now enable CONFIG_XEN. > > There is also zero reason to do this. > > When the XEN code or whatever setups up the bridge, LRO will be > disabled by the bridge code. > > It automatically issues ethtool LRO disable commands to the devices > involved in the bridge configuration. And it has done this for > a _long_ time. The ipv4 routing code will do the same. > > This change is bogus and likely based upon behavior in some ancient > distribution kernel that lacks the ethtool facilities. > > I absolutely refuse to apply this. Agreed. It is often devices that implement LRO, but don't have way to turn it off. -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/drivers/net/bnx2x_main.c b/drivers/net/bnx2x_main.c index a35def6..ab9a9eb 100644 --- a/drivers/net/bnx2x_main.c +++ b/drivers/net/bnx2x_main.c @@ -52,6 +52,10 @@ #include <linux/stringify.h> +#if defined(CONFIG_XEN) +#define __NO_TPA__ 1 +#endif + #include "bnx2x.h" #include "bnx2x_init.h" #include "bnx2x_init_ops.h"